- From: Howard Katz <howardk@fatdog.com>
- Date: Tue, 20 Jul 2004 14:11:53 -0700
- To: "Rob Shearer" <Rob.Shearer@networkinference.com>, "Steve Harris" <S.W.Harris@ecs.soton.ac.uk>, <public-rdf-dawg@w3.org>
Hmm, I'm much more partial to that solution now than when you first presented it to me. Not sure why. I was probably so focused on the deep ineffable beauty :-) of my own design that I wasn't that open to other possibilities. Or maybe I was thinking you only accessed the XML serialization, which isn't all that interesting. In any event I need to mull some more ... Howard > The idea is that 'asserted()' is in fact an external function; in the > simplest case we're assuming the RDF data is just sitting in the context > somehow, not necessarily accessible in any way other than these external > functions. > > > -----Original Message----- > > From: Howard Katz [mailto:howardk@fatdog.com] > > Sent: Tuesday, July 20, 2004 10:20 AM > > To: Rob Shearer; Steve Harris; public-rdf-dawg@w3.org > > Subject: RE: Proposed XQuery requirement and/or objective > > > > Steve, > > > > Just to clarify Rob's example a bit (you don't mind, do you, Rob?): > > > > 1) Rob's using an XML document with one or more <Person> > > elements in it. It > > looks like this could be any sort of XML document, not necessarily an > > RDF/XML serialization. (Is that correct, Rob?) > > > > 1) His use of "doc(http://foo/people.xml)/Person" implies a > > single document > > with a single <Person> as the root element. Since that's not > > very useful, > > it's more likely he meant to say: > > > > doc( http://foo/people.xml" )//Person > > > > which implies that the document has mulitple <Person> > > elements in it, the > > more likely scenario; the descendant operator (//) then > > dereferences off the > > single document node returned by doc() and grabs all its > > <Person> elements > > at one fell swoop. > > > > 2) If it's straightforward XML and didn't have any xmlns: namespace > > declaration against <Person>, you don't need a namespace > > prefix to retrieve > > it via XPath. > > > > 3) Whether the step operator is "/" or "//", neither usage implies an > > rdf:type. Rob's working in XML and he's getting XML <Person> > > elements back, > > not RDF. > > > > Rob: > > Just to clarify something, when you say: > > > > asserted( $member/URI, ...#worksFor, ...#NetworkInference ) > > > > is asserted() an external function (ie, one that's allowed to > > manipulate > > *anything*, not necessarily just items that are in the XQuery > > data model) > > that has access to an actual RDF graph or graphs and is doing > > "real" RDF > > querying against their triples directly? That would make > > sense to me if so. > > > > Howard > > > > > -----Original Message----- > > > From: public-rdf-dawg-request@w3.org > > > [mailto:public-rdf-dawg-request@w3.org]On Behalf Of Rob Shearer > > > Sent: Tuesday, July 20, 2004 9:18 AM > > > To: Steve Harris; public-rdf-dawg@w3.org > > > Subject: RE: Proposed XQuery requirement and/or objective > > > > > > > > > > > > > > for $member in doc(http://foo/people.xml)/Person > > > > > where asserted($member/URI, http://foo#worksFor, > > > > http://foo#NetworkInference) > > > > > return $member/name > > > > > > > > Wouldn't "Person" in the first line require some > > > > namespace/prefix? Also, I > > > > dont quite understand why the / after doc() implies the > > > > rdf:type part of the > > > > ($memeber rdf:type foo:Person) > > > > triple. Is it just shorthand? if so, why "/"? > > > > > > No; this example was meant to demonstrate use of XML along with RDF > > > stuff, so the XPath expression was just selecting all the top-level > > > <Person> elements in some XML doc in the conventional way. > > Everything > > > but the asserted() call is vanilla XQuery. > > > > > > >
Received on Tuesday, 20 July 2004 17:10:33 UTC